We all read and hear about different and often contradictory medical research results. If you still watch a traditional TV, the one which was created last century and carry this burden into this one, you will find an example of a medical research discovering something new almost instantly. And no, the results are never very consistent, which they actually should be if we were talking about real science. I am not trying to say that medical research is not science. Far from it. These people do a wonderful and important work saving lives and finding solutions to some terrible medical conditions.
However, there are some aspects of medicine and life in general, which medical science should not study the way it does now. The subjects I am talking about are all related to truly or perceived harmful substances or nutritional elements. You just cannot study them with such approach. Think about this for a moment. How should a well-designed study should look like? Our best and standard practice would be a blind or even better double blind, placebo controlled variety. What does that mean practically? Simple. We divide our subject into two groups. One gets the substance we are about to investigate another one is a control group, which gets a placebo. By the way, if you wonder what blind and double blind means, it is easy too. A blind study is one where the subjects do not know to which group they belong and what substance they are getting. Double blind means that even a researcher who dispatches the substance and placebo to both groups does not know who gets what, meaning probably that he does not know who belongs to which group.
Now we are getting to a juicy part of all this. Let’s take my favorite example. A common claim that smoking or ingesting marijuana causes brain damage and IQ reduction in teenagers. We can start with discussing the very weak link between IQ and any damage to a brain. Obviously, the way we measure IQ plays here a very significant role. If an IQ is a measurement of how well our prefrontal cortex does problem-solving, which it generally is then what if taking marijuana simply changes how other parts of the brain operate. For example, if it redistributed blood toward other centers, which are responsible for creativity. This aspect of our brain activity is not measured by standard IQ tests. But we went on a tangent here for a moment. Let’s get back to our proposed study. Pay attention. This is important. We propose to take a group of teenagers with a similar IQ indicators. We will divide them into two groups and start giving them marijuana and placebo. They will not know who gets what. By the end of this study, we hope to cause a detectable brain damage in the group who is getting the marijuana. Please, sign here if you agree to take part in our well-planned investigation. Obviously absurd. Ethically, you cannot do that and besides, you will have no subjects.
And this is true with all such studies. Meat cause cancer. Vegetarian diet makes you ill and vitamin deficient. Eating genetically modified food makes you genetically modified. You can make your own list when you are bored. It does not matter what is the subject. What matters is an underlying principle. Scientifically, it is impossible to study causing harm you can only study positive effects of methods and substances. Which I find wonderful, comforting and uplifting. We are so talented in making things go bad that we really can use less of it. OK, how then we produce so many medical researches of this real or perceived harm? Easy again. And also explains why we are not getting any consistent results. We use what is called epidemiological studies. It means that you gather information from the subject about their habits, what they eat, how much and so on or in the case of marijuana how often they smoke and other details. Then you analyzed your data against your chosen substance. And then you get correlation at best.
There is a very important scientific and almost philosophical principle of “correlation does not mean causation”. And this brings us to a pinnacle of our analysis of this problem. To prove causation you must cause something. Any other methods prove only correlation. Therefore, you need to cause harm in order to prove it. Makes a meaningful harm studies impossible to do. So, next time you read that a news study finds that substance X cause Y, remember what does that actually mean.